Homepage Solution manuals Ivan Niven An Introduction to the Theory of Numbers Exercise 5.7.21* (The unbounded component of $\mathscr{C}_f(\mathbb{R})$ is a subgroup of index $2$)

Exercise 5.7.21* (The unbounded component of $\mathscr{C}_f(\mathbb{R})$ is a subgroup of index $2$)

Let 𝒞 f ( ) be defined by (5.50) where a , b , c are real numbers, and suppose that the the polynomial on the right side of (5.50) has three real roots r 1 < r 2 < r 3 . Let 𝒞 0 be the connected component of points ( x , y ) 𝒞 f ( ) for which x r 3 , including the point at infinity, and let 𝒞 1 denote the connected component of points for which r 1 x r 2 . Let P and Q be arbitrary points of 𝒞 f ( ) . Show that P + Q lies on 𝒞 0 , or 𝒞 1 , according as P and Q lie on the same, or different, components (That is, 𝒞 0 is a subgroup of index 2 in 𝒞 f ( ) .)

Answers

Proof. Let 𝒞 f ( ) be defined by y 2 = p ( x ) , where

p ( x ) = x 3 + a x 2 + b x + c = ( x r 1 ) ( x r 2 ) ( x r 3 ) , ( r 1 , r 2 , r 3 ) 3 , r 1 < r 2 < r 3 .

(thus a = r 1 + r 2 + r 3 ).

Then 𝒞 f ( ) is the union of the graph of h and of the graph of h , where

h { [ r 1 , r 2 ] [ r 3 , + [ x p ( x ) = x 3 + a x 2 + b x + c = ( x r 1 ) ( x r 2 ) ( x r 3 ) .

Let 𝒞 0 be the connected component of points ( x , y ) 𝒞 f ( ) for which x r 3 , including the point at infinity, and let 𝒞 1 denote the connected component of points for which r 1 x r 2 . Since h is continuous, h ( [ a , b ] ) is bounded, so 𝒞 1 is the bounded connected component, and 𝒞 0 is the unbounded component of 𝒞 f ( ) .

Let P = ( x 1 , y 1 ) , Q = ( x 2 , y 2 ) be arbitrary points of 𝒞 f ( ) . We assume without loss of generality that x 1 x 2 . If P + Q = ( x 3 , y 3 ) , by the explicit formulas (5.52), (5.53), we have

x 3 = λ 2 a x 1 x 2 ,

where λ = y 2 y 1 x 2 x 1 if P Q , and λ = p ( x 1 ) 2 y 1 if P = Q .

Suppose first that P and Q are on 𝒞 1 , so that r 1 x 1 x 2 r 2 . It is not easy to prove directly that x 3 r 3 , so we prove this fact indirectly.

Lemma 1. Every (projective) line cuts the unbounded component 𝒞 0 .

Proof of Lemma 1. Let L be any line of the plane. If L is a vertical line, with equation x = k , and homogeneous equation x = k t , then the point at infinity O = ( 0 : 1 : 0 ) is at the intersection of L and 𝒞 0 . We suppose now that L has equation y = m x + p , and we show that there exists x 0 r 3 such that f ( x 0 , m x 0 + p ) = 0 .

Consider the function φ : defined by φ ( x ) = f ( x , m x + p ) . Then

φ ( r 3 ) = f ( r 3 , m r 3 + p ) = ( m r 3 + p ) 2 0 .

Since φ ( x ) = ( m x + p ) 2 ( x r 1 ) ( x r 2 ) ( x r 3 ) + x 3 ,

lim x + φ ( x ) = ,

so there is some r such that φ ( r ) < 0 .

Since φ is continuous on , and φ ( r 3 ) 0 , φ ( r ) < 0 , by the Intermediate Value Theorem, there is some x 0 r 3 such that φ ( x 0 ) = 0 , so f ( x 0 , m x 0 + p ) = 0 . Then the point M 0 = ( x 0 , m x 0 + p ) L 𝒞 0 . □

Now consider the line L passing by the points P = ( x 1 , y 1 ) and Q = ( x 2 , y 2 ) on 𝒞 1 ( L is the tangent at P if P = Q ). By the Lemma, there is a point R = ( x 3 , y 3 ) L 𝒞 0 . But a line has exactly three intersection points with the elliptic curve 𝒞 f ( ) (counting multiplicity), so 𝒞 f ( ) L = { P , Q , R } . This shows that P + Q + R = 0 , thus P + Q = R 𝒞 0 . This shows that

( P 𝒞 1  and  Q 𝒞 1 ) P + Q 𝒞 0 . (1)

Lemma 2. Let A = ( r 1 , 0 ) 𝒞 1 . For every point M 𝒞 0 , there is a unique N 𝒞 1 such that M = A + N .

Proof of Lemma 2. Consider the point A = ( X 1 , Y 1 ) = ( r 1 , 0 ) 𝒞 1 . Then 2 A = O . Consider now any point M = ( X 2 , Y 2 ) 𝒞 0 , and N = A + M = ( X 3 , Y 3 ) . We show first that N 𝒞 1 . By the explicit formula (5.52), using a = r 1 + r 2 + r 3 and X 1 = r 1 ,

X 3 = ( Y 2 Y 1 X 2 X 1 ) 2 a X 1 X 2 = Y 2 2 ( X 2 r 1 ) 2 + r 2 + r 3 X 2 = r 2 + Y 2 2 ( X 2 r 1 ) 2 ( X 2 r 3 ) ( X 2 r 1 ) 2 = r 2 + ( X 2 r 1 ) ( X 2 r 2 ) ( X 2 r 3 ) ( X 2 r 1 ) 2 ( X 2 r 3 ) ( X 2 r 1 ) 2 = r 2 ( r 2 r 1 ) ( X 2 r 3 ) X 2 r 1 r 2 ,

because r 1 < r 2 < r 3 X 2 . Therefore X 3 r 2 , so N = A + M 𝒞 1 , and A + N = 2 A + M = M . This shows that every point M 𝒞 0 is of the form A + N , N 𝒞 1 . Moreover N is unique, because M = A + N = A + N implies N = A + M = N .

We suppose now that P = ( x 1 , y 1 ) and Q = ( x 2 , y 2 ) are on 𝒞 0 , and we want to show that P + Q 𝒞 0 . By Lemma 2, there are points P 1 , Q 1 on 𝒞 1 such that P = A + P 1 , Q = A + Q 1 . Then P + Q = ( A + P 1 ) + ( A + Q 1 ) = 2 A + P 1 + Q 1 = P 1 + Q 1 𝒞 by (1). We have shown

( P 𝒞 0  and  Q 𝒞 0 ) P + Q 𝒞 0 . (2)

Finally, if P 𝒞 0 and Q 𝒞 1 , then by Lemma 2, P = A + P 1 , where P 1 𝒞 1 , and P + Q = A + P 1 + Q = A + R , where R = P 1 + Q 𝒞 0 by (1), thus P + Q = A + R 𝒞 by (2). So

( P 𝒞 0  and  Q 𝒞 1 ) P + Q 𝒞 1 . (3)

We have shown that P + Q lies on 𝒞 0 , or 𝒞 1 , according as P and Q lie on the same, or different, components.

Moreover, if P = ( x , y ) is on 𝒞 0 , then P = ( x , y ) 𝒞 0 . This shows with (2) that 𝒞 0 is a subgroup of 𝒞 f ( ) .

By Lemma 2 and (3), 𝒞 1 = A + 𝒞 0 , therefore

𝒞 f ( ) = 𝒞 0 ( A + 𝒞 0 ) ,

so 𝒞 0 is a subgroup of index 2 in 𝒞 f ( ) .

User profile picture
2025-06-04 09:30
Comments